headers
Geoffrey Leach | 2 Sep 2009 20:43
Gravatar

Message from long test

>From the test results:
# 1  Extended offline    Interrupted (host reset)      50%  

I did not do this deliberately. What's its saying?

Thanks.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 1 comment |
headers
Dan Lukes | 2 Sep 2009 21:46
Picon

Re: Message from long test

Geoffrey Leach napsal/wrote, On 09/02/09 20:43:
>>From the test results:
> # 1  Extended offline    Interrupted (host reset)      50%  
> 
> I did not do this deliberately. What's its saying?

Something resets the drive controller.

Computer restart may be reason, for example.

It's not easy to say who or what on your computer with unknown 
configuration and operating system triggered such reset ...

					Dan

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 0 comments |
headers
Mario 'BitKoenig' Holbe | 3 Sep 2009 10:22
Picon
Picon
Favicon

smartd device type autodetection: 5.38 vs. 5.39 r2879

Hello,

due to increasing dynamics in linux kernel's device naming, I switched
my smartd.conf from /dev/sdX names to the more stable /dev/disk/by-id/X
symlinks.
smartd 5.38 (from smartmontools_5.38-3_i386.deb) autodetects the device
type despite the symlinks. smartd 5.39 r2879 does not do this anymore
and instead logs `unable to autodetect device type' for each name. It
logs this even if the symlink does not exist at all. I now have to
explicitely set the device type via -d ata.

I'm not sure if this change is intentional - and if it is, why :)

When I use /dev/sdX, smartd 5.39 r2879 still autodetects the device
type:
	Device: /dev/sda, type changed from 'scsi' to 'sat' 
	Device: /dev/sda [SAT], opened 

regards
   Mario
--

-- 
<eckes> wenn die nsa spam filtern wuerde...
<Neral> einfach china kappen
*** Signoff: Chinah (ix.irc.at linz.irc.at)

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
(Continue reading)

Permalink | Reply | 1 comment |
headers
Alex Samorukov | 4 Sep 2009 10:52
Picon

-d sat device type on FreeBSD

Hello,

i`m still working on moving from CONTROLLER_* constants in 
os_freebsd.cpp to the new interface and now I found references about 
"sat" device type. Is it really used on BSD or is a copy-paste from 
linux? I see that it is referenced in 5.38 also, but i don`t see any 
code working with it.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 9 comments |
headers
Christian Franke | 4 Sep 2009 11:23
Picon
Favicon

Re: -d sat device type on FreeBSD

Alex Samorukov wrote:
> i`m still working on moving from CONTROLLER_* constants in 
> os_freebsd.cpp to the new interface and now I found references about 
> "sat" device type. Is it really used on BSD or is a copy-paste from 
> linux? I see that it is referenced in 5.38 also, but i don`t see any 
> code working with it.
>
>   

The "-d sat" and "-d usb*" options are handled in the OS independent 
default implementation of smart_interface::get_smart_device().

If SAT works through the default SCSI ("-d scsi") I/O control, there 
should be no need to handle "-d sat" in 
freebsd_smart_interface::get_custom_smart_device().

Cheers,
Christian

PS: Could you please put related changes (e.g. os_freebsd.cpp+CHANGELOG) 
in one single svn commit? These are easier to review then.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 1 comment |
headers
Dan Lukes | 4 Sep 2009 11:32
Picon

Re: -d sat device type on FreeBSD

Alex Samorukov napsal/wrote, On 09/04/09 10:52:
> i`m still working on moving from CONTROLLER_* constants in 
> os_freebsd.cpp to the new interface and now I found references about 
> "sat" device type. Is it really used on BSD

It returned from parse_ata_chan_dev() when user manually specified type 
(-d sat).

The return code is then used in freebsd_smart_device::open() where mean 
"known controller that doesn't require controller-type specific handling 
here".

You can discard such constant, bud you need another constant with the 
similar meaning unless you rewrite the freebsd_smart_device::open() logic.

						Dan

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 6 comments |
headers
Alex Samorukov | 4 Sep 2009 11:44
Picon

Re: -d sat device type on FreeBSD

Dan Lukes wrote:
> Alex Samorukov napsal/wrote, On 09/04/09 10:52:
>   
>> i`m still working on moving from CONTROLLER_* constants in 
>> os_freebsd.cpp to the new interface and now I found references about 
>> "sat" device type. Is it really used on BSD
>>     
>
> It returned from parse_ata_chan_dev() when user manually specified type 
> (-d sat).
>
> The return code is then used in freebsd_smart_device::open() where mean 
> "known controller that doesn't require controller-type specific handling 
> here".
>
> You can discard such constant, bud you need another constant with the 
> similar meaning unless you rewrite the freebsd_smart_device::open() logic.
>
>   
Thanks, already found this.

 Yes, my final goal is to rewrite logic of the guess_type and open to be 
more like to os_linux.cpp. Also I did USB autodetection for the FreeBSD8 
(completely different USB stack) and commited patches from the freebsd 
port. When i will have some time i`ll try to convert existing logic to 
something more readable.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on
(Continue reading)

Permalink | Reply | 5 comments |
headers
Alex Samorukov | 4 Sep 2009 11:48
Picon

Re: -d sat device type on FreeBSD

Christian Franke wrote:
> Alex Samorukov wrote:
>> i`m still working on moving from CONTROLLER_* constants in 
>> os_freebsd.cpp to the new interface and now I found references about 
>> "sat" device type. Is it really used on BSD or is a copy-paste from 
>> linux? I see that it is referenced in 5.38 also, but i don`t see any 
>> code working with it.
>>
>>   
>
> The "-d sat" and "-d usb*" options are handled in the OS independent 
> default implementation of smart_interface::get_smart_device().
>
> If SAT works through the default SCSI ("-d scsi") I/O control, there 
> should be no need to handle "-d sat" in 
> freebsd_smart_interface::get_custom_smart_device().
Yes, found that. For now it should work correctly. BTW, I found that 
usb* device types are not exported to the help, i did ticket in the TRAC 
for this.
>
> PS: Could you please put related changes (e.g. 
> os_freebsd.cpp+CHANGELOG) in one single svn commit? These are easier 
> to review then.
Ok, i`ll do.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
(Continue reading)

Permalink | Reply | 0 comments |
headers
Dan Lukes | 4 Sep 2009 12:04
Picon

Re: -d sat device type on FreeBSD

Alex Samorukov napsal/wrote, On 09/04/09 11:44:
>>> i`m still working on moving from CONTROLLER_* constants in 
>>> os_freebsd.cpp to the new interface and now I found references about 
>>> "sat" device type.

> Yes, my final goal is to rewrite logic of the guess_type and open to be 
> more like to os_linux.cpp

I didn't monitor the progress of os_linux.cpp so much. Do Linux's gueess 
device type from /dev/* entry name ? If yes, then we shall not follow 
such way.

We can use system enumeration to found devices (and indirectly - device 
types as well) - see os_freebsd.cpp:get_dev_names_{ata|scsi}()

It's more flexible way.

> Also I did USB autodetection for the FreeBSD8 (completely different USB stack)

I hope I will found the reason why USB devices are not reachable via 
standard CAM (SCSI) layer. When patched then nobody needs distinguish 
USB devices as special case.

Also, SAT devices should be accessible as standard ATAPICAM devices, but 
it's long run way now.

> When i will have some time i`ll try to convert existing logic to 
> something more readable.

I tried it, but failed. I have no sufficient experience with
(Continue reading)

Permalink | Reply | 4 comments |
headers
Alex Samorukov | 4 Sep 2009 12:15
Picon

Re: -d sat device type on FreeBSD

Dan Lukes wrote:
> Alex Samorukov napsal/wrote, On 09/04/09 11:44:
>   
>>>> i`m still working on moving from CONTROLLER_* constants in 
>>>> os_freebsd.cpp to the new interface and now I found references about 
>>>> "sat" device type.
>>>>         
>
>   
>> Yes, my final goal is to rewrite logic of the guess_type and open to be 
>> more like to os_linux.cpp
>>     
>
> I didn't monitor the progress of os_linux.cpp so much. Do Linux's gueess 
> device type from /dev/* entry name ? If yes, then we shall not follow 
> such way.
>
>   
We already using this way in os_freebsd if there is no -d switch. More 
than this - even with -d switch we are using device names to find 3ware 
controller type (there are 3 different types, mapped to twa or twe nodes).
> We can use system enumeration to found devices (and indirectly - device 
> types as well) - see os_freebsd.cpp:get_dev_names_{ata|scsi}()
>
> It's more flexible way.
>   
I think they are currently used in the smartd.
>   
>> Also I did USB autodetection for the FreeBSD8 (completely different USB stack)
>>
(Continue reading)

Permalink | Reply | 3 comments |

Gmane